The language (and machines) the Internet is built on.

Data Warehousing Battles Are Harsh

James Curran recently found himself face to face with a data warehousing time bomb that would not be dismantled by any technology fix. The senior vice president of management information services at State Street Bank & Trust Co., in Boston, wanted to change the format for some key financial data, but the users in control were balking. Grounds for a skirmish? Not according to Curran, who instead opted for compromise to keep his data warehousing project from detonating.

After heated debate, Curran agreed to leave this particular data alone. In return, the users promised to show support for the warehouse in other ways that were important to Curran’s group. Specifically, they agreed to start using a new feature that lets users input annotative data, including notes and comments, on financial activity, helping Curran’s staff create a financial intelligence database as a core component of the system.

“You can compromise on the tactical issues,” Curran says of the bank’s data-warehousing efforts, “but don’t give up on the strategy.” State Street’s data warehouse environment comprises a back end built on Microsoft Corp.’s SQL Server database, along with Arbor Software Corp.’s Essbase multidimensional database and Pilot Software Inc.’s Lightship as front-end data-access tools.

Curran and other IS managers are realizing that the road to data warehousing is littered with as many people and political land mines as it is with technology obstacles. Turf battles, user resistance, and power struggles can be as critical to the success of a data warehouse as choosing the right back-end database or design schema. IS managers need to be properly prepared to effectively manage people and finesse organizational issues.

“There are a number of technical solutions, but the cultural issues are more difficult to overcome because if individuals see this as a threat to power, then resistance will surface,” says George Trudel, a consultant with the Business and Technology Office at Blue Cross and Blue Shield of Rhode Island. The Providence-based insurance company is using various databases and tools for its warehouse strategy, including IBM’s DB2.

Man in the middle

In essence, a data warehouse is a database loaded with data extracted from operational systems and built for the express purpose of end-user access. More so than many IS efforts, data warehousing can ignite controversy because it often spans multiple departments, which do not necessarily share the same goals. This hits on the sensitive issue of data ownership, often leaving IS in the delicate position of being in the middle. “The best shot you have, from an IS perspective, is to stay out of the middle,” says Frank McGuff, a consultant and director of strategic initiatives at Braun Technology Group, in Chicago.

To avoid this dilemma, IS managers need to put more focus on data management and data ownership. Projects have a better shot at survival when these potentially explosive issues get resolved up front. Observers say it is critical to figure out who owns the data and who is responsible for the data warehouse before any other steps are taken.

That certainty would have made things easier at one insurance company working on a data warehouse, recalls Rich Finkelstein, president of Performance Computing Inc., in Chicago, who was brought in to consult on the project. In this case, one group of users wanted the warehouse to provide summary data, while another group was looking to IS to deliver detailed data.

To keep both groups happy, Finkelstein helped the company’s IS team build a database that was “optimized for neither” user group.

A business case

Successful projects often place overall warehouse responsibility with the business unit because it best understands the business reasons for doing the warehouse in the first place, managers and consultants say. Plus, the business unit has the money to fund the project. Increasingly, business units are assuming ownership of the data by taking responsibility for what goes into the warehouse and assuring the quality of information. That’s resulted in a new role, dubbed data steward (see chart, below).

With the business units handling these chores, IS is free to carry the technical load of the project, including design and implementation of the warehouse. Several IS managers say their organizations also handle training and troubleshooting, in part because those are traditional jobs of IS.

That’s what’s happening at a large Midwest industrial equipment manufacturer. Rick Schulte, project leader for EIS and marketing systems, says the company is building a warehouse based on relational and multidimensional technologies and using data from multiple departments. IS has long maintained the sales and marketing data and will continue to do so. But, when it comes to financial data, the financial department is responsible for choosing the data and maintaining quality.

The reason? The financial staff has a history of maintaining its own data. Schulte decided to keep the status quo. Schulte’s team is responsible for the warehouse itself, but it counts on the financial group to deliver accurate data.

Other established ground rules will make projects flow more smoothly. Braun Technology’s McGuff advocates setting up a conflict resolution process–including a designated arbitrator–at the outset of a warehousing project so the team can confront trouble when it crops up.

While this tactic can ease project management, it won’t be easy to pull off, McGuff adds. “There is a no-failure-on-my-watch mentality today,” McGuff says, so it’s difficult to make a case for this approach without raising the eyebrows of upper management. “If you stand up and say, ‘There will be conflict,’ executives won’t want to hear it.” The trick, McGuff says, is to finesse a positive spin.

Companies such as State Street Bank are also looking to end-user cooperation as another essential component to a successful data warehouse. But IS managers say users can run hot and cold. At first, they may be disinterested, but when they discover what a warehouse can provide, they can become zealots.

To win them over, IS managers need to prove the warehouse has benefits and “seeing is believing,” says Jerry Silva at PacifiCare Health Systems Inc., in Cypress, Calif. Silva, who is heading up a corporatewide data-warehousing project based on an Oracle Corp. database, says, “You have to keep showing progress or else you will lose support.”

Silva’s staff kept users’ attention by providing them with an initial proof-of-concept document and then following up with a prototype last year. A beta version of the warehouse will be rolled out in April, and Silva’s staff will release new functions in June, with extras on tap for October.

But Silva and other IS managers also caution that it is essential to keep users’ expectations in line. Says Wayne Eckerson, an analyst with Patricia Seybold’s Office Computing Group in Boston: “If it isn’t scoped out, as soon as IS gets the pilot done, users will start submitting requests for all kinds of new data and new views. It can really get out of hand.”

To avoid this scenario, Eckerson suggests that IS sharpen its project-management skills and keep close tabs on the work. That way, IS can handle user requests with more authority–and talk cost. “That will make users think twice,” he adds.

Finally, IS managers advise getting a high-level executive to sponsor a data-warehouse project. Steve Goodfriend, a manager of systems and analysis at Tropicana Products Inc., in Bradenton, Fla., says his firm’s data-warehousing effort has benefited from high-level “champions,” including the company president and controller. “I have stomped my feet about projects that have to get done, but until the right person comes along and supports it, it doesn’t happen,” Goodfriend says.

posted by Coder Carl in Uncategorized and have No Comments

Place your comment

Please fill your data and comment below.
Your comment